Consider $$ U \equiv \dot{\theta} - \theta \ ,$$

where as per usual $\dot{\theta} = \frac {d \theta}{dt}$. We then have $$ \frac {\partial U}{\partial \theta} = -1 \quad , \quad \frac {\partial U}{\partial \dot{\theta}} = 1 $$

Here is what happens if we try to compute the former partial derivative in Mathematica:

Of course, one work-around is to replace $\theta$ with say $x$ but that becomes cumbersome, especially when one has a lengthy and convoluted expression.

Anyone know how to resolve this?

EDIT: What I want to achieve is that Mathematica recognizes the two different involved quantities as two fundamentally different variables. Say we had $U \equiv \stackrel{lol}{\theta} - \theta$. Then

Same issue.


You need to define OverDot[\[Theta]] to be the derivative of \[Theta] with respect to some variable t; otherwise there is no way for Mathematica to know the relation between \[Theta] and OverDot[\[Theta]].

OverDot[\[Theta][t_]] := D[\[Theta], t];


D[OverDot[\[Theta][t_]]- \[Theta], \[Theta]]



D[OverDot[\[Theta][t_]]- \[Theta], Overdot[\[Theta][t]]


  • $\begingroup$ The point is it does not necessarily have to be defined as the derivative of $\theta$. Say the expression is $\stackrel{cool}{\theta} - \theta$. The same problem arises. The point is I want Mathematica to recognize $\stackrel{cool}{\theta}$ and $\theta$ as two completely different variables. $\endgroup$
    – user23366
    Dec 25 '14 at 20:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.